Hello, if you have any need, please feel free to consult us, this is my wechat: wx91due
COM1005 Machines and Intelligence
Intelligent Agents: Lab Question Sheet
Instructions1.Complete the tasks below on your copy of this question sheet.
2.Download your completed copy of this Google Doc as a .docx or .odt file (note: NOT .pdf or .pages) and submit it to the Intelligent Agents Submission Dropbox on Blackboard.
The deadline for submission is 3pm on Thursday, November 7th.
Week 4: Simple Reflex Agents
Task 1
Copy and paste the code for your Transition Model in the box below, replacing the template.
Ensure the syntax is correct as the code will be tested during assessment.
tm = {
'cn': {
('?', '?', 'ng'): [
('f', 'nb'),
('tr', 'ce'),
('tl', 'cw')
],
('?', '?', 'g'): [
('s', 'w')
],
},
}
Task 2
Perform 30 game runs with Maisy-S, with each run on a newly randomly generated layout (excluding unwinnable layouts), using the transition model you submitted above and using the time_limit value of 1000.
Put your results in the table below:
Simple Reflex Agent: Trial Data
Run # 1 2 3 4 5 6 7 8 9 10
Steps
Outcome
Run # 11 12 13 14 15 16 17 18 19 20
Steps
Outcome
Run # 21 22 23 24 25 26 27 28 29 30
Steps
Outcome
Simple Reflex Agent: Statistics
# of games Wins Losses / Time-outs Steps:
Min Steps:
Max Steps:
Median Steps:
Mean Steps:
St. Dev.
30
Week 5: Model-Based Reflex Agents
Task 3
Perform 30 game runs with the Model-Based Reflex Agent (Maisy-B), adhering to the following rules:
a.Each run should be done on a new, randomly generated layout.
Use the new_game functionality to generate new layouts, but make sure to exclude unwinnable maze layouts.
b.Use the same Transition Model and the same value of the time_limit you have used in the previous lab for the trials with the Simple Reflex Agent (Maisy-S).
Put the results in the tables below.
Model-Based Reflex Agent: Trial Data
Run # 1 2 3 4 5 6 7 8 9 10
Steps
Outcome
Run # 11 12 13 14 15 16 17 18 19 20
Steps
Outcome
Run # 21 22 23 24 25 26 27 28 29 30
Steps
Outcome
Model-Based Reflex Agent: Statistics
# of games Wins Losses / Time-outs Steps:
Min Steps:
Max Steps:
Median Steps:
Mean Steps:
St. Dev.
30
Task 4
Using the information in the maze_layout.md file and save_maze and load_maze functionality, create two maze layouts:
(a) one where the MBRA, on average, reaches the goal cell in fewer steps than SRA, and
(b) another one where the SRA, on average, reaches the goal cell in fewer steps than MBRA.
Paste the screenshots with the corresponding maze layouts, followed by their binary representations in the table below:
a.
b.
For each case, provide the average number of steps and explain why you think one agent outperformed the other.
ANSWER [100 words]:
Task 5
Considering the setting of this particular game (“A-Maisy Game”), indicate which agent model is better suited for the task and explain why.
ANSWER [200 words]: